The scientific study of language in the 20th Century was concerned primarily with uncovering systematic regularities in sounds and their sequences. This is true both of the structuralist approaches of the first half of the century, and of the generative enterprise that dominated the second half. Both schools sought properties of human communication that admitted of highly compact symbolic description, were universal, and were sufficiently abstract to work in the domains of both spoken and written communication. The intellectual context of the birth of generative linguistics was enriched by the advent of a mathematical theory of information and communication (Shannon, 1948), the emerging view of mind as a form of symbolic processing supported by neurons (McCulloch, & Pitts, 1943), and the birth of a cognitive psychology that saw syntactic regularities at the heart of human cognition (Fodor, 1975). The emphasis on universal, abstract, and symbolic processes necessarily led to a sharp distinction between the linguistic and nonlinguistic features of human languaging (i.e. all the many behaviors in evidence when humans communicate). In this view, the roles of speaker and listener are utterly distinct, and the medium of communication—voice or writing—is passed over. This has become a thoroughly conventional view of what language is: Language is characterized as a form of message passing, from one unseen mental domain to another. The message-passing view has facilitated the development of many technologies, from writing itself, to the modern forms of mass textual proliferation. 

The commitment to a view of language as message passing, and the concentration on the universal, the abstract, and the symbolic, has as a consequence that many aspects of human languaging are ignored, to the point of rendering them invisible. Indeed, the focus on the kind of regularities studied by mainstream linguistics could be seen as the study of the symbolic structure of writing, rather than speaking/listening, for it is only with writing, and widespread literacy, that the very idea arises of words that have regular repeatable structure, but that are not vouched for by an utterer. To speak a sentence is an act with consequences. That act is not included in the traditional linguistic account. Writing and literacy have been argued to induce wholesale and significant cognitive changes in their users (Ong, 1982; Olson, 1996). Older by far than writing is the voice and vocal behavior. When vocal interaction takes place it unfolds in real-time among mutually engaged participants, who speak from concerned perspectives, and for whom the act of uttering is, itself, meaning creating. There are rich interactions between the voice, “body language”, gaze, etc. The roles of speaker and listener are often not entirely distinct. Vocal interactions are often not sensibly interpretable as the passing of messages, or encoded propositions, at all. In the focus on langue over parole, and in the valorization of competence over performance, linguistics seems to have lost its voice! 

The term “Joint Speech” is introduced as an umbrella label for all those situations in which many people say the same thing at the same time. Specific forms of joint speech are found in ritualized practices of collective prayer in all major religions. Joint speech is regularly found in situations of protest and demonstration. Unison chanting is found among sports fans the world over. Choral speaking is a performative variety of joint speaking especially popular in South East Asia. Public oaths of allegiance are typically executed collectively. The takbir is a type of collective expression of shared belief very common in the Islamic world. Joint speaking practices find specific niche applications in educational practices, including memorization and pronunciation training. 

As the above list makes clear, situations in which we speak as one are both ubiquitous, and they are all characterized by a great degree of collective intentionality and purpose. Put simply, they are situations that we care deeply about. It is therefore perplexing that there does not exist a canon of scientific work on joint speech. By my own informal reckoning, there are more than ten times as many scholarly articles on the marginal phenomenon of glossolalia, or speaking in tongues, as there is on joint speech. Yet there are clearly very many scientific questions that arise in the context of joint speech: What is its form? What purposes does it serve? What regularities are common across these situations, and what are the domain-specific characteristics? What is the relation between the words and the actions of the speakers? What can we infer from a situation in which joint speaking arises? This list could be 

extended many times over, and yet we seem to have developed a blind spot for joint speech. Why is this? 

I will argue that the orthodox view of language as message passing from one mind to another leads to this blindness for a suite of phenomena that are of prima facie importance to all humans. In its commitment to a thoroughly Cartesian metaphysics, it establishes a view of language that illuminates those characteristics shared by the voice and writing, but that thereby hides the richness and plurality of human vocal behavior. In losing the voice, linguistics has managed to ignore the many other descendants of voice. They are found in ritual, rite, and liturgy. They are found in music and song (Cummins, 2013). There may be many of them. 

I will discuss joint speech from several perspectives. Among these is the laboratory study of synchronous speech, in which subjects are asked to read a novel text in synchrony with one another (Cummins, 2003). This experimental vehicle has brought several empirical characteristics of joint speech to light. In interpreting such findings, it will be necessary to approach human languaging from a non-Cartesian point of view. The technical tools of dynamical systems theory (Kelso, 1995), and the philosophical framework of enaction (Stuart et al., 2010; Cummins, 2013a) seem to provide an appropriate formal vocabulary for describing many aspect of joint speech. Speech errors that arise during joint speaking appear to sometimes be of a kind never found otherwise. In many cases, an error by one speaker will cause the abrupt and simultaneous cessation of speaking in both speakers. This points to the need to describe the coupling among speakers as an important empirical feature, and it suggests that speaker/listeners may become interlinked in collective speech in ways that have manifest consequences. Preliminary evidence of a marked alteration to cortical activity when speaking in unison with another live person, but not when speaking in unison with a tape recording, will also be presented. Although early, a large empirical agenda arises. 

The study of joint speech is in its infancy, but the field seems to be enormously rich and varied, and to be integrated into very many important domains of human activity. I thus dare to hope that the presentation of the work at the Dinafon meeting might encourage some participants to turn their inquisitive gaze towards some of the phenomena that thereby arise, but that have passed beneath the scientific radar due to the manner in which language has conventionally been considered.